scales 1.0.0

We’re delighted to announce the release of scales 1.0.0. The scales packages provides the internal scaling infrastructure to ggplot2 and its functions allow users to customize the transformations, breaks, guides and palettes used in visualizations in ggplot2 and beyond.

This is a major release with significant changes to the popular formatter functions, and added transformations. Below we demonstrate some of the biggest changes in this release including:

  • updated formatter functions
  • new transformations
  • improved breaks on log-transformed scales

See the News for a detailed list of changes and visit the new website for complete documentation and additional usage examples.

Install the latest version with:


Formatter changes

Thanks to the help of @larmarange, scales has now added a generic formatter function, number_format(), that powers the majority of the formatters in the package.

This change altered the arguments of most of the formatters: comma_format(), percent_format(), unit_format(), dollar_format(), ordinal_format(), and scientific_format(). All gained new arguments accuracy, scale, prefix, and suffix, allowing users to specify rounding accuracy, a scaling value, a prefix, and a suffix to customize output. Furthermore, all of these functions now allow user specification of the thousands separator, big.mark, and the decimal separator, decimal.mark. Users of these formatters will notice that the default thousands separator is now a space. This default was chosen as a useful compromise for an international audience (a space for thousands separator and a dot for decimal separator) and is officially endorsed by SI/ISO 31-0 standard, as well as by the International Bureau of Weights and Measures and the International Union of Pure and Applied Chemistry (IUPAC), the American Medical Association’s widely followed AMA Manual of Style, and the Metrication Board, among others (read more here). Those interested in number formatting with no separator can simply define big.mark = "" to revert to the previous behaviour.


number(c(12.3, 4, 12345.789, 0.0002))
#> [1] "12"     "4"      "12 346" "0"

# these functions round by default, but you can set the accuracy
number(c(12.3, 4, 12345.789, 0.0002), big.mark = "", accuracy = .01)
#> [1] "12.30"    "4.00"     "12345.79" "0.00"

# percent() function takes a numeric and does your division and labelling for you
percent(c(0.1, 1 / 3, 0.56))
#> [1] "10.0%" "33.3%" "56.0%"

# comma() adds commas into large numbers for easier readability
#> [1] "10,000,000"

# dollar() adds currency symbols speficifed by `prefix` or `suffix`
dollar(c(100, 125, 3000))
#> [1] "$100"   "$125"   "$3,000"
dollar(c(100, 125, 3000), suffix = "€", prefix = "")
#> [1] "100€"   "125€"   "3,000€"

# unit_format() adds unique units
# the scale argument can do simple conversion on the fly
unit_format(unit = "ha", scale = 1e-4)(c(10e6, 10e4, 8e3))
#> [1] "1 000 ha" "10 ha"    "1 ha"

Three additional formatters have been added: pvalue_format() formats p-values, number_bytes_format() formats numeric vectors into byte measurements, and time_format() provides support for formatting POSIXt and hms objects. Finally, ordinal_format() has gained new rules for French and Spanish formatting.

New transformations

Two new transformations were added to the package for this release: psuedo_log_trans(), and modulus_trans(). pseudo_log_trans() transforms data on a signed logarithmic scale with a smooth transition to a linear scale around 0. The modulus_trans() was added along with a refactored boxcox_trans() to provide a better option for negative numbers. Both modulus_trans() and boxcox_trans() gained an argument offset which now allows users to fit both type-1 and type-2 Box-Cox transformations.

Better breaks

In a long awaited fix, log_breaks() now returns integer multiples of integer powers of base when finer breaks are needed on the log scale. This will change all ggplot graphics with log-transformed axes.


dsamp <- sample_n(1000, tbl = diamonds)
ggplot(dsamp, aes(y = price, x = carat)) + 
  geom_point() + scale_x_log10() + scale_y_log10()


We’re grateful to the 24 people who contributed issues, code and comments: @alexandreliborio, @AndreaCirilloAC, @batpigandme, @BenOnEarth, @billdenney, @Bisaloo, @CesarSancho, @christianhomberg, @clauswilke, @cwickham, @dpseidel, @foo-bar-baz-qux, @graciecorgi, @hadley, @jimhester, @jnolis, @larmarange, @lepennec, @markvanderloo, @ptoche, @RobertMyles, @statist7, @ThierryO, and @zeehio.

Upcoming events
Bellevue WA
May 29 - June 1
Mara Averick, Garrett Grolemund, Javier Luraschi, Max Kuhn, and Kevin Kuo will be teaching workshops on text mining, the tidyverse, deep learning, and tidy modeling.
Toulouse, France
July 9
Jim Hester, Hadley Wickham, and Jenny Bryan are teaching a half-day tutorial on Package Development.